基于改进AdaTAD的奶山羊爬跨行为时序动作定位方法OA
Temporal Action Localization of Mounting Behavior in Dairy Goats Based on Improved AdaTAD
[目的/意义]奶山羊爬跨行为的时序定位是繁殖管理的重要基础.针对现有方法多停留在行为判别层面、在未修剪视频中对短时突发行为的起止边界刻画不足,且易受遮挡、视角变化与背景干扰影响的问题,提出一种基于面向时序动作定位的适配器调优(Adapter Tuning for Temporal Action Detection,AdaTAD)改进的端到端时序动作定位方法,以实现爬跨行为的准确识别与起止时间精确定位.[方法]以AdaTAD框架为基线,引入视觉提示调优,通过少量可学习Prompt Tokens对主干注意力分布进行任务引导,增强关键帧及边界邻域的特征响应;设计多尺度运动感知适配器,采用并联多尺度时序深度可分离卷积分支建模不同时间尺度的运动模式,并结合残差连接与非线性映射稳定注入主干特征,提升短时微动作与相对完整动作过程的联合建模能力.[结果和讨论]所提方法的平均精度均值达到81.72%,相较基准模型AdaTAD提升5.00个百分点;在时间交并比为0.7的更严格条件下达到68.85%,较AdaTAD提升4.06个百分点,表明该方法在高边界精度要求下仍具有优势.模型推理速度为65.78帧/s,可训练参数量为27.941 M,在精度提升的同时保持较低开销.[结论]该方法可提升复杂养殖场景下奶山羊爬跨行为的时序定位精度与稳定性,为繁殖行为监测与管理决策提供关键时序信息支撑.
[Objective]Accurate temporal localization of mounting behavior in dairy goats is important for intelligent reproductive man-agement,as event frequency,onset time,and duration provide useful evidence for heat monitoring and mating decisions.Unlike sim-ple behavior recognition,temporal localization in untrimmed videos enables fine-grained,time-resolved records for practical farm use.However,real-world mounting behavior is usually brief and sporadic,with few informative frames in long video streams.Moreover,weak discrimination from similar non-target interactions,together with occlusion,viewpoint variation,and background motion,often degrades boundary-aware representation learning and leads to unstable start-end localization.To enhance localization accuracy and sta-bility while maintaining practical efficiency for deployment,an improved AdaTAD-based end-to-end temporal action localization ap-proach was proposed for mounting behavior in dairy goats.[Methods]The proposed approach adopted AdaTAD as the baseline end-to-end temporal action localization framework and introduced two complementary improvements,explicit key-frame guidance and multi-scale motion modelling,while retaining the original detection head and post-processing pipeline for generating temporal action in-stances.First,visual prompt tuning(VPT)was incorporated to provide task-conditioned guidance to backbone feature extraction in a parameter-efficient manner.Specifically,a small number of learnable prompt tokens were inserted into the Transformer backbone with backbone parameters frozen.Through multi-head attention interactions between prompt tokens and patch tokens,the prompts steered attention towards mounting-relevant temporal regions,strengthened feature responses at critical frames and in boundary neighbour-hoods,and improved the separability between brief target segments and abundant background frames.Second,a multi-scale motion adapter(MSMA)was introduced to model motion patterns at different temporal scales and improve robustness to diverse scene dy-namics.MSMA employed parallel multi-scale temporal depthwise separable convolution branches to capture short-,mid-,and longer-range temporal variations,enhancing representations of subtle short-duration micro-actions as well as relatively complete action pro-cesses.Residual connections and nonlinear mappings further stabilised feature injection and gradient propagation,enabling multi-scale dynamics to be integrated into backbone features with limited additional optimisation burden.Overall,VPT focused on bound-ary-relevant attention guidance,whereas MSMA emphasised multi-scale temporal dynamics modelling;Together,they formed a com-plementary design within the end-to-end localization pipeline.[Results and Discussions]Comparative experiments showed that the pro-posed method achieves an average mAP(mean Average Precision@[0.3:0.1:0.7])of 81.72%,improving upon the baseline AdaTAD by 5.00 percentage points,indicating that incorporating VPT and MSMA enhanced overall localization performance.At a temporal Inter-section over Union(tIoU)threshold of 0.7,the proposed method attained 68.85%,exceeding AdaTAD by 4.06 percentage points,dem-onstrating that the performance gain was preserved under stricter temporal boundary-consistency criteria.Further comparisons with representative approaches,including TadTR,VSGN,AFSD,ActionFormer,TriDet,DyFADet,and Re2TAL,showed average mAP im-provements of 38.82,33.83,25.29,4.09,2.83,1.20,and 6.06 percentage points,respectively,demonstrating stronger overall competi-tiveness.In terms of efficiency,the model ran at 65.78 f/s with 27.941 million trainable parameters,indicating that the accuracy gains were achieved while maintaining a relatively low parameter overhead and practical runtime efficiency.Overall,task-guided prompting and multi-scale temporal modelling improved key temporal feature representations with limited parameter increments,thereby benefit-ing localization of short,sporadic behaviors.[Conclusions]By combining VPT for boundary-relevant attention guidance with an MS-MA for multi-scale temporal dynamics modelling,the proposed AdaTAD-based end-to-end temporal action localization method im-proves localization accuracy and maintains stable advantages under stricter boundary-consistency requirements,while preserving prac-tical inference efficiency.The method could provide temporal information for reproductive behavior monitoring and decision support,and offer a feasible basis for building individual-level,time-resolved management systems in real farming environments.
王佳源;李其同;罗元滔;杨蜀秦;王振华;宁纪锋;王美丽
西北农林科技大学信息工程学院,陕西 杨凌 712100,中国西北农林科技大学信息工程学院,陕西 杨凌 712100,中国西北农林科技大学信息工程学院,陕西 杨凌 712100,中国西北农林科技大学机械与电子工程学院,陕西 杨凌 712100,中国西北农林科技大学信息工程学院,陕西 杨凌 712100,中国西北农林科技大学信息工程学院,陕西 杨凌 712100,中国西北农林科技大学信息工程学院,陕西 杨凌 712100,中国
农业科技
爬跨行为时序动作定位AdaTAD精准养殖多尺度运动建模端到端模型智慧养殖
mounting behaviortemporal action localizationAdaTADprecision livestock farmingmulti-scale motion modelingend-to-end modelsmart breeding
《智慧农业(中英文)》 2026 (4)
148-163,16
国家重点研发计划项目(2022YFD1300200)陕西秦创原引用高层次创新创业人才项目(QCYRCXM-2022-359) National Key Research and Development Program of China(2022YFD1300200)Shaanxi Qinchuangyuan High-lev-el Innovation and Entrepreneurship Talent Program(QCYRCXM-2022-359)
评论